Transit
Optimizing trunk bus lines could improve service
Metrobus riders, especially those in the suburbs, often face long wait times between buses. With WMATA's budget crisis, it is extremely unlikely they can add more bus service. However, there are ways WMATA could dramatically improve service at little to no cost.
Photo by Beechwood Photography.
Other cities, starting with Portland's Tri-Met, have had great success with their bus networks by creating high-frequency lines. In these instances, bus riders don't need to carry a schedule because the bus shows up too often for that to be necessary. This has the benefit of increasing convenience and mobility for riders and, consequently, ridership.
Many Metrobus routes have common segments with other bus lines, yet because they are different routes, they are often not scheduled with each other in mind. One example of this case can be seen in Metrobus routes C2, R12, and T16/17 in Greenbelt. Each of these routes serves several common points, including Greenbelt Metro, Beltway Plaza, and Greenbelt Center. There are variations in the routes, but there are also many common segments.
Metrobus C2, R12, and T16/17 routes in Greenbelt. Click to enlarge.
The C2 operates every 15-20 minutes during rush hours and every 30 minutes for most of the rest of the day. The R12 and T16/17 each operate at 30 minute headways during rush hours and hour intervals otherwise.
This could easily translate into relatively high frequencies, even off peak. With four vehicles an hour, each of the common points could see a vehicle every 15 minutes. Headways would be even better during peak periods. However, that's not how WMATA operates these buses. During the midday period, for example, these buses depart Greenbelt Station in the direction of Greenbelt Center at 1:23 (R12), 1:24 (C2), and 1:25 (T16). The last bus, a C2, departed 29 minutes prior, and the next bus, a C2, won't depart for 28 minutes. For passengers headed toward Greenbelt, there's no reason that the time between the 12:54 C2 and the 1:53 C2 (59 minutes) can't be divided into 20 minute waits with evenly spaced departures of the buses.
The above chart shows a section of a weekday. The top gray bar indicates scheduled bus departure times in the direction of Greenbelt Metro from the timepoint "Greenbelt Center", located at the corner of Crescent Road and Gardenway. The bottom bar shows departures from Greenbelt Metro in the direction of Greenbelt Center. Notice on the lower bar that departure times on the three routes from Greenbelt Metro are clustered in roughly 3-5 minute intervals at about 25 after the hour each hour. Offsetting the T16/17 and R12 by several minutes would even out departure times, lowering average wait times for those bound for any of the common route segments.
The similar departure times for the C2, R12, and T16/17 are typical throughout the day, although during rush hours, higher frequencies help to even things up a bit. On trips toward Greenbelt Metro, it is typical for the R12 and T16/17 buses to be immediately behind one another during their entire common run through Old Greenbelt. For passengers, this situation is intolerable - to wait 30 minutes to see two buses bound for the same destination following each other. During "bunching" situations, this is upsetting. For it to be a scheduled, everyday occurrence, is completely unacceptable.
Rescheduling buses on trunk lines is one way to increase service, at least as far as the customer sees it, at very little cost to the agency. Technically, there's no increase in capacity through rescheduling - the same number of buses, the same number of seats - but a better distribution of those seats throughout the day.
In some instances, it might make sense to reroute neighborhood buses to serve a particular trunk or to give more common segments. Simplification of route structure could increase productivity by shortening the length of routes and making it easier for riders to understand the bus line.
For instance, routing the T16/17 along the same route as the R12 between northern Greenbelt and Greenbelt Metro would decrease travel time to the Metro for Greenbelters. Extending T16/17 buses to a terminal at Beltway Plaza would maintain current service levels to the mall. It would also balance service levels between Greenbelt Center and Greenbelt Metro. Right now, the C2, with the same frequency as the combined R12/T16/17, travels via Beltway Plaza between Greenbelt Center and the Metro. The R12 makes that trip via the northern reaches of Greenbelt. The T16/17, however, travels via both before getting to the Metro.
Potential bus route modifications. Click to enlarge
These ideas are just simple ways that WMATA could cut costs while improving service. There are probably other instances where these solutions could be applied throughout the region. From a rider's point of view, headway improvements are too long in coming. In these lean times, these adjustments would be a valuable approach.
 | Matt Johnson has lived in the Washington region since mid-2007. He has a Master's degree in Community Planning from the University of Maryland and a BS in Public Policy from Georgia Tech. He has worked in the planning field since 2006 and lives in Greenbelt, where he serves on the city's Advisory Planning Board.    |
Comments
- Successful speed cameras require fair speed limits
- Amid scandal, don't lose sight of Gray's policy achievements
- VDOT ignores own data, pushes widening I-66
- Montgomery plans 160-mile, "gold standard" BRT system
- DC's divide need not be black and white
- Preservationists ask to shrink 3rd Church replacement
- Planners are the new public health officials
by Redline SOS on Feb 17, 2010 11:03 am
by Martin on Feb 17, 2010 11:08 am
by Brad on Feb 17, 2010 11:26 am
by Jasper on Feb 17, 2010 11:41 am
by Dave on Feb 17, 2010 11:53 am
by Matt Glazewski on Feb 17, 2010 11:55 am
Commenters can complain that this is a "no-brainer" and WMATA are "morons" for not doing this, but you have to take into account the fact that these routes are all different lengths, so they take different amounts of time to complete. That means the more you want things to line up nicely, the more buses you are going to need in order to make those headways match up. A bus can't magically leave the beginning of a route if it's still on the route coming the opposite way.
by MLD on Feb 17, 2010 12:04 pm
by Martin on Feb 17, 2010 12:24 pm
by Erica on Feb 17, 2010 12:37 pm
If Metro didn't change the route of the bus, didn't change the run time or distance, and only moved each the departure up by 20 minutes on one route (say the R12), and back by 20 minutes on another route (say the T16/17), no additional cost would be incurred. But, intervals would be better on common segments - at least the ones analyzed for this post.
You are correct that Metro couldn't just move one of the dots. They'd have to move all the yellow dots by the same amount in the same direction to keep costs from increasing.
In the graphic above, better headway spacing could be achieved by moving all the yellow dots (at Greenbelt Metro, for instance) 20 minutes to the left and all the blue dots 20 minutes to the right.
That's really all I'm calling for in terms of scheduling.
It's also true that that change might not make for even spacing at other timepoints, but it could make for better spacing. Especially if the planning takes that into account.
Now, I did propose some potential route adjustments. Those would either shorten routes or be pretty close to no net gain on route miles. So that would actually tend toward a decrease in the number of vehicles (and operators) required to run the route. Of course, savings would have to pass a certain threshold before a bus could actually be removed from the schedule.
You are correct that it would not be easy to schedule every bus line that had common segments with other buses so that intervals were nice and even. But what's more important? Convenience to transit planners or convenience to transit riders? Any effort would be a plus on these routes.
@everybody else:
Note that I didn't call Metro's planners morons. They're hard working, intelligent people. They have hard jobs and they're short-staffed. And without feedback from the community, they don't always know that improvements we value most.
And more importantly, many of them read this blog. So if you're going to criticize, at least be constructive. They might actually listen.
by Matt Johnson on Feb 17, 2010 1:36 pm
WMATA would serve more passengers -- and better serve them -- if it focused more on regular, on-time service along core routes, optimizing stop placement and frequency to the point that people with a choice would, in fact, choose to use the bus. Right now the mission of serving those most in need, while noble, ties up too many resources and still fails to serve those needy customers well.
Focusing on core routes could also allow WMATA to rearrange feeder and shuttle service, allowing those who do live farther from the core to rely more on the on-time performance they could expect from shorter routes. Take a short feeder route to the terminus of a core route, then get on whichever core bus happens to show up next. Such a system would also reduce the negative effect of a single missed or delayed route, since shorter feeders could run more frequently, and a more frequently served core would provide more travel options.
by fedward on Feb 17, 2010 2:00 pm
No, this is not a one-person operation. In fact, we have about 25 active contributors or so. You can read all about us here (though not everyone has a profile):
http://greatergreaterwashington.org/about.cgi
David Alpert is the founder and primary contributor. He serves on Metro's Riders Advisory Committee as a representative of the District of Columbia.
You can discern who writes each post in the byline. Clicking on the author's name will take you to all the posts he or she has written.
by Matt Johnson on Feb 17, 2010 2:04 pm
And, MoCo Ride-On has the same bunching problem, on lines originating/going to the same terminal. Had it for years
by KenF on Feb 17, 2010 2:50 pm
Scheduling is an NP-hard problem, which means that you are extremely unlikely to find an optimal solution in a reasonable period of time, witness the problems that schools have had with scheduling. However, plenty of algorithms and software exist to find near-optimal solutions. It just takes some skill in programming and operations research to use them. It's long been clear that Metro has near-zero OR expertise. Metro could contract out the scheduling. Metro can also simplify its route structure to make scheduling easier.
Before Metro even attempts this, it needs to collect good data on ridership system-wide. Segments with low ridership can be dropped or have service reduced.
Metro also needs to think creatively, which right now seems oxymoronic. I had suggested to no avail adding the G2 to the D-series study as it covers roughly the same route. As it is, the eastbound G2 has to make a left turn to reach P St., which is a time-killer.
by Chuck Coleman on Feb 17, 2010 6:38 pm
1. Identify important core routes, including looking at ways to reduce proximal, parallel routing. Looking at Georgia Avenue and 11th, 14th, and 16th Streets for instance, it might be worth picking just two, assuming frequency is more valuable than convenience.
2. Assign a global percentage of vehicles to the core as a whole; reassign vehicles from that pool to maintain average headways across routes. The design goal here has to be headways, because the system is more valuable to riders the less they stand around waiting for the next bus. In fact, reducing standing around waiting should be a design goal for Metrorail as well -- I used to have to wait 19 minutes one night a week on my trip home to Crystal City from Shady Grove. I hated it but that was the best possible route.
3. Optimize shorter feeder routes based on remaining equipment.
4. Where possible reduce the effect of traffic on core routes, with stop light priority, dedicated lanes, smarter routing, and other low-overhead improvements as available.
On the last point, for instance, the Georgia Avenue route would be much improved if the buses didn't have to stop nearly every block. The 79 bus is what, eight minutes faster than the 70 from one end to the other? How much better would that be with optimized stop lights?
The same is true on 14th Street: the southbound Circulator is great up until it gets to Thomas Circle, at which point you can sometimes expect it to take another fifteen or twenty minutes to get you to McPherson Square. Maybe the optimized 14th Street route should just turn around at the circle, with downtown distributor routes running on the other side.
by fedward on Feb 17, 2010 8:37 pm
For example, I was waiting for a 30's bus to come by my stop on Wisconsin Ave. this evening (off-peak). I waited for over 15 minutes for a 36 bus to come, but right behind it was a 32 and a 31 bus. I've seen this occur before, when a bus is packed, but the one right behind it is almost empty... talk about a waste of money and fuel.
If Metro did a little tweaking of the bus schedules to make it so that buses on overlapping lines would be more evenly spaced, bus riders' experiences would be drastically different.
by John on Feb 18, 2010 12:59 am
The new trolley helps a lot, but it still means a 15-20 minute wait to go down King at some points and is very, very crowded in many instances.
by Lisse on Feb 18, 2010 3:19 pm
Add a Comment